Chapter 17
Coming from the Mouths of Babes
In This Chapter
Tracking children’s speech patterns
Distinguishing healthy and disordered speech processes
Applying this knowledge for transcription
Adults aren’t the only people you’ll transcribe in your phonetics classes and in your real-world career. For anyone working in speech language pathology, understanding child language is a must. The same holds true for anyone interested in the fields of childhood education, child language research, or dialectology. In this chapter, I take you through the periods of (healthy) speech development, discuss key differences between healthy and disordered speech, and give you some tips on how to put this knowledge into practice in your transcriptions.
Following the Stages of a Healthy Child’s Speech Development
Knowing how children’s speech develops is an important part of phonetics. Here you can track the sounds produced by children from the age of 6 months to 2 years old. I highlight universal aspects of young children’s speech production and touch on some of the theories proposed to account for these amazing aspects of children’s behavior.
Focusing on early sounds — 6 months
The first sounds to come out of a young infant are shaped by the physical capabilities of that very young person. When you’re only a few months old, you don’t have much of an adult-like vocal tract. The larynx is high in the throat and only begins to descend to adult-like proportions at approximately 5 to 7 months. At this stage in a person’s life, these sounds are pretty much limited to high-pitched squeals, grunts, and cries.
Nevertheless, children at this age engage in a remarkable amount of communication, despite the inability to form words. They communicate with gaze, by imitating the pitch of their caretakers’ speech, by making facial expressions, and by gesturing.
Babbling — 1 year
By approximately one year of age (often starting around 9 months), children begin the phase known as babbling, producing short, repeated utterances. This behavior, much beloved by parents, plays a major role in infant-parent bonding behavior.
Babbling is broadly described as having two phases:
Reduplicative: This term refers to repeated speech. An example of reduplicative babbling would be “ba-ba-ba-ba” or “goo-goo-goo.”
Variegated: This term refers to many different sounds. Variegated babble consists of longer strings and more varied sounds than reduplicated babble. Some researchers also describe a jargon phase (occurring at about 10 months of age) at which adult-like stress and intonation begin to kick in. An example of variegated babbling would be “ka-be-to-gi-ta-ge.”
Children babble when they’re relaxed and comfortable. This behavior is thought to be a way of engaging the yet-developing vocal folds. Early babbling isn’t necessarily related to communication, although babbling carries over into early word production.
Speech babble has provided researchers valuable insights into infant behavior. For instance, the rhythmic opening and closing gestures of children’s mouths in forming utterances such as “buh-buh” and “ga-ga” have been interpreted in the Frame-Content Theory. This theory teases out the rhythmic opening and closing (syllabic) part of infant babbling behavior (called the Frame) from the segment-specific elements (such as consonants and vowels), called the Content. According to this view, a babbled syllable isn’t a random mix of consonants and vowels, but instead motoric constraints result in the following pairs:
Alveolar consonant and front vowel (such as /di/ and/de/)
Labial consonant and central vowel (such as /bʌ/ and /ba/)
Velar consonant and back vowel (such as /ɡo/ and /ɡu/)
So far, researchers have found such patterns in English-speaking infants and in child speakers of other languages (including Swedish, Japanese, Quechua, Brazilian-Portuguese, Italian, and Serbian). These findings have spurred on other researchers to investigate to what degree babbling is shaped by the growth of the vocal tract itself versus other developmental processes, such as the maturation of the motor control system (or the need for infants to first discover and then fine-tune relationships between their speech movements and sounds).
Researchers have also found that young children open the right side of their mouths more when they babble, suggesting that the left side of the brain controls this babbling.
Due to the physiological limits of young children, some sounds tend to be produced more than others. A study of 15 different languages, including English, Thai, Japanese, Arabic, Hindi, and Mayan, showed the following consonants commonly occur:
/p/, /b/, /m/, /t/, /d/, /n/, /s/, /h/, /w/, /j/
However, these phonemes were rarely found:
/f/, /v/, /θ/, /ð/, /ʃ/, /ʒ/, /ʧ/, /ʤ/, /l/, /ɹ/, /ŋ/
These data suggest that early babbling is at least partly independent of language-particular factors.
Forming early words — 18 months
Hearing a child’s first words is one of the most rewarding experiences of being a parent. For a phonetician, studying the sound patterns in those first words is just about as exciting.
Young children can hear sound contrasts well before they can produce them. Just because they have immature articulatory systems doesn’t mean that their sharp little minds aren’t doing well at teasing out the sounds big people are telling them.
In terms of what children want to say, the most common items in the first 50 words are typically nouns, including such words as “daddy,” “mommy,” “juice,” “milk,” “dog,” “duck,” “car,” “book,” and “blocks.” Young children follow with verbs and adjectives, including properties (“all gone,” “more,” and “dirty”), actions (“up,” “down,” “eat,” “seat,” and “go”), and personal-social terms (“hi,” “bye,” “please,” and “thank you”). By the time children have acquired 50 words or so (usually by around 18 months of age), they start to adopt fairly regular patterns of pronunciation.
Although children vary a good deal in terms of the order in which they master speech sounds in production and perception, the following general tendencies seem to exist:
As a group, vowels are generally acquired before consonants (by age three).
Stops tend to be acquired before other consonants.
In terms of place of articulation, labials are often acquired first, followed (with some variation) by alveolars, velars, and alveo-palatals. Interdentals (such as /θ/ and /ð/) are acquired last.
New phonemic contrasts occur first in word-initial position. Thus, the /p/ to /b/ contrast, for instance, shows up in pairs such as “pat” and “bat” before “cap” and “cab.”
Toddling and talking — 2 years
A two-year-old is a very different creature than a six-month old. The motoric and cognitive systems are much further developed (and, true, they generally relish saying “no!”). This section describes the sound inventory you can expect in English for a two-year-old talker.
By age 2, a typical English-speaking child has the following inventory of consonant phonemes:
Oral stops: /p/, /t/, /k/, /b/, /d/, and /g/
Nasals: /m/ and /n/
Fricatives: /f/ and /s/
Approximants: /w/
Still to be acquired are the interdental fricatives (/θ/ and /ð/) and the voiced alveo-palatal fricative (/ʒ/). These sounds are typically acquired after age 4.
In general, the relative order in which children acquire sounds reflects the sound’s distribution in the world’s languages. The sounds that are acquired early tend to be found in more languages, whereas the sounds that are acquired late tend to be less common across languages.
Knowing What to Expect
Everyone knows that, compared to adults, children make mistakes in their speech. However, determining whether a child’s speech is healthy or disordered isn’t as easy. Because children acquire speech structures over time, certain errors are expected at certain ages. These normal (healthy) patterns of development can be contrasted with disordered child language processes.
A basic way to start thinking about whether a child’s speech is disordered (and a question familiar to many parents) is to ask: What sounds should my child be saying at such-and-such age? When answering this question, clinicians consider children’s phonological processes when evaluating healthy and disordered patterns of development, which I explain in the following sections.
Eyeing the common phonological errors
Phonologists begin by studying the errors that healthy children make when learning language. These data show many commonalities across languages, including languages from very different language families. Phoneticians generally agree that children’s phonological errors include the following:
Boo-boos at the level of syllable production
Substitutions of one consonant or vowel segment for another of like kind
Assimilation processes, in which one sound becomes more like one another
Table 17-1 gives you some examples:
Table 17-1 Common Childhood Errors
|
Syllable-Level Processes |
Example |
Production (IPA) |
|
Weak syllable deletion |
“potato” |
/ˈtedo/ |
|
Final consonant deletion |
“book” |
/bʊ/ |
|
Reduplication |
“baby” |
/bibi/ |
|
Cluster reduction |
“climb” |
/kaɪm/ |
|
Substitutions |
||
|
Stopping |
“soup” |
/tup/ |
|
Fronting |
“cake” |
/tek/ |
|
Deaffrication |
“jump” |
/ʒʌmp/ |
|
Liquid gliding |
“like” |
/waɪk/ |
|
Vocalization(liquid becomes vowel) |
“line” |
/jajn/ |
|
Assimilatory Processes |
||
|
Labial |
“pot” |
/pɑp/ |
|
Alveolar |
“mine” |
/naɪn/ |
|
Velar |
“harden” |
/ˈhargn̩/ |
|
Prevocalic voicing |
“tap” |
/dӕp/ |
|
Devoicing |
“ride” |
/ɹaɪt/ |
This table contains examples that probably seem familiar or even cute to the average person. For example, saying /ˈtedo/ for “potato.” An adult may create these kinds of errors when trying to imitate child speech.
Examining patterns more typical of children with phonological disorders
Child language specialists also seek to determine patterns that can serve as a warning of phonological disorders in children. Experts differ somewhat on the best ways to classify these disorders; however, they generally agree on the types of underlying problems. Two key concepts include
Certain children may have a language delay by showing persisting normal processes that last longer than they are supposed to.
Some children show unusual, idiosyncratic, or atypical deviance in the application of phonological rules, compared to other children.
Table 17-2 shows some examples of idiosyncratic phonological processes in child language:
Table 17-2 Errors Suggesting Childhood Speech Disorders
|
Disorder |
Example |
Production (IPA) |
|
Glottal replacement |
“stick” |
/stɪʔ/ |
|
“better” |
/ˈbɛʔɛ/ |
|
|
Backing |
“test” |
/kɛst/ |
|
“smash” |
/smӕg/ |
|
|
Initial consonant deletion |
“guess” |
/ɛs/ |
|
“kiss” |
/ɪs/ |
|
|
Stops replacing a glide |
“yellow” |
/ ˈdɛdo/ |
|
“wait” |
/bet/ |
|
|
Fricatives replacing a stop |
“quit “ |
/kwɪs/ |
|
“duck” |
/zʌk/ |
These idiosyncratic cases wouldn’t likely be included in the average adult’s imitation of child speech. The typical parent probably wouldn’t always know what is normal and what is worrisome, hence why he or she should seek a professional opinion.
Children with recognizable speech errors may have the following disorders:
Speech sound disorders: These disorders include both articulatory errors and problems with phonological development.
Childhood apraxia of speech: A motor speech disorder in which children know what they want to say but have difficulty mapping these intended sounds into realized speech movements.
Dysarthria: A motor speech disorder involving problems with the muscles of the mouth, face, or respiratory system.
Orofacial myofunctional disorders: Also known as tongue thrust, these disorders involve an exaggerated protrusion of the tongue during speech and/or swallowing.
Stuttering: A fluency problem marked by disruptions in the production of speech sounds that can impede communication.
Voice disorders: They include problems in producing sound at the level of the larynx.
For more information about these different disorders, contact the following organizations:
www.asha.org/public/speech/disorders/childsandl.htm (United States)
www.caslpa.ca/ (Canada)
www.rcslt.org/ (United Kingdom)
www.asha.org/members/international/intl_assoc.htm (Other countries, from Argentina to Vietnam)
Transcribing Infants and Children: Tips of the Trade
The exact reasons why you’re transcribing can guide you in the tools to use and in the way you do your transcription. If you’re creating transcriptions (from recordings) for clinical or teaching purposes, then you have many possible options to choose from. For example, you can be more or less narrow (transcribing fine-grained detail), incorporate certain characters from the ExtIPA (extensions of the IPA), and use a variety of different conventions to represent prosody (melody) — (see Chapters 10 and 11 for more information).
However, if you’re working in a lab or clinic that has an established protocol, you need to master those specific tools. In this section, I introduce you to a variety of methods and techniques that can be useful. I also provide you some brief examples to get you started. I include speech from the period of early word acquisition (9 to 16 months). In addition to these examples of healthy speech, I also provide a snippet of speech from a 2-year-old child with a cochlear implant to show how speech presents as children adapt to prosthetic hearing.
Delving into diacritics
In a perfect world, cleanly articulating children would produce only lovely substitution errors for your corpus. You would then transcribe little Jimmy’s production of /fɪs/ for fish, consider it a backing error (see Table 17-1), and feel darn good about yourself.
However children’s actual speech is far messier. There are errors both at the phonemic (such as substitutions, or metathesis, the switching of sounds) and phonetic (for instance distortions and coarticulatory) levels. You typically need to complete a systematic narrow transcription, indicating allophonic variation of individual phonemes. This usually requires the use of several diacritics, marks to fine-tune transcription. I introduce diacritics in Chapter 3 and further describe them in Chapter 19.
Table 17-3 lists diacritics useful for working with children’s speech, sorted by voicing, place, and manner of articulation.
Here are some more practical tips for when transcribing children’s speech:
Don’t become frustrated. You can’t be expected to identify every phoneme your talker produces.
Circle the features you do know, work on the rest later. For instance, if you know the phoneme is a voiced fricative, you can write:
Take frequent breaks. Don’t listen to a sound more than three times in a row.
Keep your mind clear and don’t read into the transcription what is not there. I have seen many transcriptions that reflect what the transcriber thought (or desperately hoped) would come next.
Study No. 1: Transcribing a child’s beginning words
The first sample is from a project performed by professor Marilyn Vihman at Stanford University (currently at the University of York, England), investigating the beginning of children’s phonological organization. Table 17-4 shows transcriptions that come from a young child babbling as she approached her first words.
These transcriptions include parentheses for sounds produced quietly (hə) and (ə), and light aspiration is shown with a superscript “h” in parentheses. Vowel lengthening (using the diacritic [ː]), glottal stop, and nasalization are noted. A question mark after “[kɛt]” indicates the transcriber was unsure of this transcription.
Study No. 2: A child with a cochlear implant (CI)
The second study performed by Andrea Warner-Czyz, PhD, at the University of Texas at Dallas, includes data from a young girl, H, profoundly deaf from birth, fitted with a cochlear implant (CI), activated when she was 11 months 22 days old. This girl was considered a successful CI user. The following minitable shows some utterances transcribed 13 and 18 months post-implant.
|
Time Post Implant |
Parent |
Child Response (IPA) |
|
13 months post |
Mommy, see the baby. |
/ma mi ʃi ə be bi/ |
|
18 month post |
Pick him up |
/i jə bəp/ |
|
18 month post |
Hey, Mommy. Sit down. |
/eː mami di do/ |
The key purpose of these data was to identify basic errors (at the phonemic level) and to track the expansion of the child’s phoneme repertoire. As such, the researcher conducted a fairly broad transcription. Features such as vowel length were detailed, using [ː] for long vowels and elsewhere [ːː] for extra-long vowels. Patterns of omission/substitution/metathesis were described, and unexpected patterns of intonation are indicated. In most cases, phonetic departure from targets is indicated with substituted IPA symbols (for instance, /ʃ/ found for the /s/ target of “see” in the line 1).
Babbling birds: Clues to speech mysteries
Birds babble. Not just in cartoon commercials, but they also babble when young songbirds are acquiring their songs. It depends on the exact species and how they’re exposed to their adult songs. Here scientists start to see possible connections to human behavior.
The Zebra finch is a particularly well-studied species. Young male finches first have an auditory learning period in which they must hear examples of the songs they’re to sing. They then produce a variety of immature songs called subsong. The birds then advance to the plastic stage (where some of the adult forms are noticeable), followed by mature song.
Subsong seems to be like human babble in several ways:
For both babies and birds, youngsters are attracted to their own species more than to others. Human babies are more interested in human voices than other noises, and baby birds are far more tuned in to the songs of their own species than the songs of other types of birds.
These phases of birdsong learning seem parallel to humans having a sensitive period for language learning (when language learning is best accomplished, after which point it becomes more difficult and learners are left with a non-native accent).
Like humans, if song patterns are reinforced with positive social feedback, they’re more likely to recur, which is especially prominent in social species such as the Zebra finch.
Auditory feedback plays an important role in maintaining both speech and birdsong, although researchers don’t fully understand the processes involved.
Scientists are exploring other interesting commonalities between bird communication and human speech, including similar neural bases and genetic contributions (including the FOXP2 gene). Birdsong has syllables, dialects, and accents. Scientists can learn a lot by studying feathered friends. Birdbrain isn’t necessarily an insult.
To hear song sparrows learn to sing, go to http://birdnote.org/show/song-sparrows-learn-sing.